ZK Developer's Reference/UI Composing/Richlet

A richlet is a small Java program that composes a user interface in Java for serving user's request.

When a user requests the content of an URL, the ZK Loader checks if the resource of the specified URL is a ZUML page or a richlet. If it is a ZUML page, then the ZK Loader creates components automatically based on the ZUML page's content as we described in the previous chapters.

If the resource is a richlet, the ZK Loader hands over the processing to the richlet. What and how to create components are all handled by the richlet. In other words, it is the developer's job to create all necessary components programmatically in response to the request.

The choice between ZUML pages and richlets depends on your preference. However, the performance shall not be a concern since the parsing of ZUML is optimized.

Implement a Richlet

It is straightforward to implement a richlet. First, implement the Richlet interface and then mapping URL to the richlet.

Implement a Richlet as a Java class

A richlet must implement the Richlet interface. However, you generally don't have to implement it from scratch. Rather, you could extends from GenericRichlet. With GenericRichlet, the only thing you have to do is to implement Richlet.service(Page). It is called when an associated URL is requested. For example,

 package org.zkoss.zkdemo;

 import org.zkoss.zk.ui.Page;
 import org.zkoss.zk.ui.GenericRichlet;
 import org.zkoss.zk.ui.event.*;
 import org.zkoss.zul.*;

 public class TestRichlet extends GenericRichlet {
     //Richlet//
     public void service(Page page) {
         page.setTitle("Richlet Test");

         final Window w = new Window("Richlet Test", "normal", false);
         new Label("Hello World!").setParent(w);
         final Label l = new Label();
         l.setParent(w);

         final Button b = new Button("Change");
         b.addEventListener(Events.ON_CLICK,
             new EventListener() {
                 int count;
                 public void onEvent(Event evt) {
                     l.setValue("" + ++count);
                 }
             });
         b.setParent(w);

         w.setPage(page);
     }
 }

As shown above, we have to invoke Component.setPage(Page page) explicitly to attach a root component to a page such that it will be available at the client.

To have better control, you can implement the Richlet.init(RichletConfig) and Richlet.destroy() methods to initialize and to destroy the resources required by the richlet when it is loaded.

In additions, you could implement Richlet.getLanguageDefinition() to use a different language as default (for example, implementing a richlet for mobile devices). By default, ZUL (aka., xul/html) is assumed.

Richlet Must Be Thread-Safe

Like a servlet, a single instance of richlet is created and shared for all users for all requests for the mapped URL. A richlet must handle concurrent requests, and be careful to syncrhronize access to shread resources. In other words, a richlet (the implementation of the service method) must be thread-safe.

Don't Share Components in Different Requests

When a request (not Ajax request but regular HTTP request) is made by an user, a Desktop and a Page are created first, and then Richlet.service(Page) is invoked to serve the request. In other words, each request is served with an individual desktop and page. Therefore, we cannot share components among different invocation of Richlet.service(Page).

For example, the following code is illegal:

public class MyRichlet extends GenericRichlet {
    private Window main; //Not a good idea to share
    public void service(Page page) {
        if (main == null) {
            main = new Window();
        }
        main.setPage(main); //ERROR! Causes an exception if the same URL is requested twice!
...

Why? Each deskop shall have its own set of component instances. When the URL associated MyRichlet is requested a second time, an exception will be thrown because the main window is already instantiated and associated with the first desktop created from the first request. We cannot assign it to the second desktop.

---

Map URL to a Richlet

To map URL to a richlet, there are two steps.

1. Turn on the support of Richlet (in WEB-INF/web.xml)
2. Map URL pattern to Richlet (in WEB-INF/zk.xml)

Turn on Richlet

By default, richlets are disabled. To enable them, add the following declaration to WEB-INF/web.xml. Once enabled, you can add as many as richlets as you want without modifying web.xml any more.

<servlet-mapping>
    <servlet-name>zkLoader</servlet-name>
    <url-pattern>/zk/*</url-pattern>
</servlet-mapping>

where you can use replace /zk/* to any pattern you like, such as /do/*. Notice that you cannot map it to an extension (such as *.do) since it will be considered as a ZUML page (rather than a richlet).

Map URL pattern to Richlet

For each richlet you implement, you can define it in WEB-INF/zk.xml with the statement similar to the following:

<richlet>
    <richlet-name>Test</richlet-name><!-- your preferred name -->
    <richlet-class>org.zkoss.zkdemo.TestRichlet</richlet-class><!-- your class name, of course -->
</richlet>

After defining a richlet, you can map it to any number of URLs using the richlet-mapping element as shown below.

<richlet-mapping>
    <richlet-name>Test</richlet-name>
    <url-pattern>/test</url-pattern>
</richlet-mapping>
<richlet-mapping>
    <richlet-name>Test</richlet-name>
    <url-pattern>/some/more/*</url-pattern>
</richlet-mapping>

Then, you can visit http://localhost:8080/PROJECT_NAME/zk/test to request the richlet.

The URL specified in the url-pattern element must start with /. If the URI ends with /*, then it is matched to all request with the same prefix. To retrieve the request's actual URL, you can check the value returned by the getRequestPath method of the current page.

 public void service(Page page) {
     if ("/some/more/hi".equals(page.getRequestPath()) {
         ...
     }
 }

---

Tip: By specifying /* as the url-pattern, you can map all unmatched URLs to your richlet.

Version History

Last Update : 2010/11/8

Version	Date	Content